Abstract
Background:
This study aims to construct a reliable diagnostic model for coronary artery disease (CAD) patients and explore its potential mechanism by consensus molecular subtypes of ferroptosis-related genes.
Methods:
GSE12288 and GSE20680 were downloaded from Gene Expression Omnibus database. CAD patients were divided into different molecular subtypes according to the expression level of ferroptosis-related genes. Then, the distribution of differentially expressed genes, functional annotations and immune infiltration cells between the two subtypes were compared. Finally, a prognostic model of ferroptosis-related genes in CAD was constructed and verified.
Results:
Two different molecular subtypes of CAD were obtained according to the expression level of ferroptosis-related genes. Then, a total of 1944 differentially expressed genes (DEGs) were found, among which, 236 genes were up-regulated and 1708 genes were down-regulated. In addition, 43 DEGs were ferroptosis-related genes. Functional enrichment analysis showed that these DEGs between two subtypes of CAD were mainly enriched in immune-related pathways and processes, such as T cell receptor, mTOR, NOD-like receptor and Toll-like receptor signaling pathways. We also found that 21 immune cells were significantly changed between two subtypes of CAD. The LASSO method was performed to identify and construct the 16 ferroptosis-related genes-based diagnostic signature. Diagnostic efficiency of diagnostic signature measured by AUC in the training set and validation cohort was 0.971 and 0.899, respectively.
Conclusions:
This study contributes to a more comprehensive understanding of the mechanism of ferroptosis-related genes in CAD.
Keywords:
Coronary artery disease; Ferroptosis; Molecular subtypes; Non-negative matrix factorization (NMF).